Every score comes with the model's reasoning on the paper page.
"ScienceBuddy: Recursive-in-Recursive Self-Improvement for Interactive Scientific Agents" scored 4.5/10 predicted impact on @kurateorg — Clarity 7.5 · Difficulty 6.0 · Translational 6.0